Size effects in the processing of thin metal sheets

Transcription

1 Journal of Materials Processing Technology ) 44±48 Size effects in the processing of thin metal sheets L.V. Raulea *, A.M. Goijaerts, L.E. Govaert, F.P.T. Baaijens Netherlands Institute for Metals Research, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands Abstract The present investigation addresses the in uence of grain sizes on the processing of thin metal sheets. The problem was approached in two different ways: rstly, by reduction of the sheet thickness at a constant grain size investigated in tension), and secondly, by changing the grain size at a constant sheet thickness investigated in bending). The material investigated was soft aluminum sheet, with thickness ranging from 0.17 to 2 mm. As the thickness of the specimens was chosen relatively high, geometric similarity with thin sheet materials was obtained by increasing the grain size by recrystallization. In this way the grain sizes ranging from to 600 mm 2 could be obtained. The results show that in both the experiments, the yield strength as well as the maximum load decrease with a decreasing number of grains over the thickness. For grain sizes larger than the specimen thickness, the value of the yield strength appears to increase with the grain size, whereas a strong increase of the variation is observed. The effect of this variation of material properties on the processing of these materials is demonstrated in a planar blanking process. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Grain size; Size effects; Thin sheets; Uniaxial tension; Bending; Blanking 1. Introduction With the ongoing miniaturization in electronics, there is a growing demand for the development of accurate forming processes for very thin sheets 10±40 mm). One of the problems that is encountered in the development of such processes is the variation in product quality and in the magnitude of the process force. The occurrence of this problem is directly related to the ratio of grain size to sheet thickness. As the sheet thickness decreases to the same order of magnitude as the grain size, the mechanical properties of the individual grains will dominate the properties of the sheet. The strong anisotropy of the mechanical properties of the grains, combined with the variation in orientation of the individual grains, will lead to a strong variation in the forming properties of the sheet. Another effect of the reduction of sheet thickness is the in uence of free surface, which was pointed out by Kals et al. [1,2]. In the presence of a free surface, a grain is less constrained and able to deform at a substantially lower apparent ow stress than in the bulk. With a reduction of the absolute sheet thickness, the surface zone becomes relatively more important, leading to a stronger decrease * Corresponding author. Tel.: ; fax: address: L.V.Raulea@tue.nl L.V. Raulea). in process forces than may be expected on grounds of geometric similarity. Kals et al. demonstrated the surface effect on CuZn15 with a grain size of 25 mm, as did Miyazaki et al. [3] on plate and rod-like specimens of Cu±13 at.% Al. However, Kals et al. also showed that the free surface effect is not observed if the grain size is increased to 110 mm. In the latter case, the process forces were found to increase with decreasing sheet thickness. In the present investigation, it is attempted to clarify these observations. Therefore, the occurrence of these size effects is studied in three ways: rstly, the effect of a thickness reduction at a constant grain size is studied in uniaxial extension; secondly, the effect of grain size variation at a constant specimen thickness is studied in three-point bending. Thirdly, the strong in uence of grain size on deformation behavior of thin sheets is also found in blanking experiments. Kals et al. showed that the occurrence of irregularities of blank product fracture zone and burr) seems to be connected to the limited formability of few grains in the very small blanking deformation zone. Still, they found that some blank product parameters does not seem to signi cantly change, e.g. the proportions of areas of rollover zone and smoothly sheared surface. In this investigation, the in uence of grain size in blanking of metal sheet is studied in its extreme: blanking of polycrystalline and single crystal specimens /01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S )

2 L.V. Raulea et al. / Journal of Materials Processing Technology ) 44± Experimental 2.1. Material and specimen preparation The material used in our investigation was Aluminum 99.0±99.5% 2S Half-Hard DIN 1747 Euro-norm. The asreceived material was subjected to different treatments according to the requirements of the experiment. In the case of uniaxial tension tests, we kept the grain size constant and varied the thickness, for the case of bending experiments, we kept the specimen thickness constant and varied the grain size, and for blanking experiments, the specimen was enlarged up to specimen width. For the tension specimens, the following procedure was applied: 1. The as-received sheet was annealed at 6008C for 1 h. 2. The material was rolled from 2.0 to 1.10, 0.72, 0.30, and 0.17 mm, respectively, and recrystallized at 6008C for 1 h. For all the specimens this led to a grain size of mm 2. For the bending specimens, the following procedure was used to vary the grain size: 1. The as-received sheet was annealed at 6008C for 1 h. 2. The specimens were then plastically deformed, and recrystallized at 6008C for 1 h. The grain sizes obtained were , 0.062, 0.25, 1, 5, 40, and 600 mm 2 and depended on the applied plastic deformation. All tension, bending, and blanking specimens were cooled down in the air after recrystallization at 6008C. For the blanking specimens, the same recrystallization procedure was used to enlarge the as-received material grain sizes up to single grain in the blanking deformation zone Techniques Grain sizes smaller than 1 mm 2 were estimated according to the standard method ASTM E112 and those larger than 1mm 2 were estimated by calculating the average area of the grains located within the bending deformation zone. In all the cases, the grain area was estimated on the free surface of the specimen thickness. The experiments were performed on a universal tensile testing machine. Bending experiments were performed with the three-point bending setup presented in Fig. 1. The dimensions of bending specimens were 15 mm 45 mm 1 mm. Uniaxial tensile tests were performed with the specimens presented in Fig. 2. In all the experiments, tension, bending, and blanking, the testing speed was 10 mm/min. In the tension tests, the yield strength was de ned as the 0.2% offset yield stress and the tensile strength as the maximum engineering stress. For bending experiments, the yield strength was calculated with the formula s y ˆ 3F yl b 0 s 2 1) 0 Fig. 1. Three-point bending setup dimensions in mm). where F y is the bending force at the yield point, l the free length 20 mm), b 0 the specimen width 15 mm), and s 0 the initial specimen thickness 1 mm). In order to compare the results of uniaxial tension tests and bending experiments, a thickness-to-grain-size ratio is introduced: j ˆ p s 2) A where j is the thickness-to-grain-size ratio, s the specimen thickness, A the average grain size on the specimen surface. This ratio j is an indication for the number of grains over the specimen thickness. For j < 1 there is a single grain and for j > 1 there are multiple grains over the specimen thickness. Planar blanking experiments were performed on a setup described previously [4]. The blanking specimen measured 10 mm 50 mm 1 mm. During the blanking experiments, the punch load as a function of the punch displacement as well as the displacement eld in the deformation zone were measured. To determine the latter, a camera recorded successive images of the lateral specimen side on the deformation zone see Fig. 3). During the image post-processing, a grid was applied on the surface and a technique referred to as digital image correlation was used to calculate the grid displacement eld [4]. This displacement eld can be subsequently used to determine the equivalent logarithmic strain eld [4]. Fig. 2. Tension specimen dimensions in mm). Fig. 3. The planar blanking sample.

3 46 L.V. Raulea et al. / Journal of Materials Processing Technology ) 44±48 3. Results and discussion 3.1. Uniaxial tension test: effect of varying thickness at constant grain size The effect of sheet thickness on the results of a uniaxial tension test for a grain size of mm 2 is shown in Fig. 4. It is clear that both the yield and tensile strength decrease with decreasing sheet thickness. This is in full agreement with the results of Kals et al. [1,2] who found similar results for CuZn15 with a grain size of 25 mm note: they used a length scale; in the present investigation an area scale is used). As mentioned in section 1, they attributed this effect to the in uence of the free surface on the ow stress of grains located at the specimen surface. The presence of a free surface strongly reduces the constraints on the grains near the surface, and consequently, the grains will deform at a substantial apparent lower ow stress. To study the in uence of the specimen free surface in uence, Kals et al. [1] introduced a parameter representing the ratio of the area of surface grains to the total area of the specimen cross-section: a ˆ 1 w 0 2L s 0 2L 3) w 0 s 0 where w 0 is the initial specimen width, s 0 the initial specimen thickness, and L the grain dimension. As it is shown in Fig. 5, a relative increase of the area of surface grains leads to a decrease in the yield and tension strength. In other words, the behavior of the grains located at the free surface becomes more and more dominant Bending experiments: effect of grain size at constant thickness The results of bending experiments on specimens of 1 mm thickness with various grain sizes are shown in Fig. 6. One can see that the results show a parabolic trend for the yield strength. In order to rationalize this behavior, it is convenient to distinguish between two regions: a) the region at the left of the graph Fig. 6) with multiple grains over the specimen Fig. 5. Variation of tensile and yield strength with a constant grain size mm 2 ). thickness, and b) the region at the right of the graph Fig. 6) with only a single grain over the specimen thickness. In the case of multiple grains over the thickness, the yield strength is observed to increase with decreasing grain size. This effect of grain size is well known and usually referred to as grain-size strengthening or the Hall±Petch effect. The grain boundaries act as barriers for dislocations, and therefore a relative increase of the grain boundary surface which occurs with the reduction of grain size will lead to an increase of the yield strength. Interesting is also the effect observed on the right-hand side of the graph for the two yield strength curves, obtained from two different experimental approaches. Firstly, the bending yield strength curve, which was obtained by varying the grain size, and secondly, the tension yield strength curve, which was obtained by decreasing the specimen thickness. It is clear that both the experiments show the same trend: the yield strength increases with an increase of the thicknessto-grain-size ratio j. Obviously the effect of an increase j by decreasing the grain size leads qualitatively to the same effect as an increase of j by increasing the sheet thickness. This observation gives rise to an idea that both the Hall± Petch effect [5] and the free surface effect [1±3] might have a similar origin. In the case of a single grain over the thickness of the sheet left-hand side of Fig. 7), two effects are observed: a) Fig. 4. Variation of tensile and yield strength with sheet thickness constant grain size mm 2 ). Fig. 6. Variation of maximum load and force at yield point for bending experiments constant specimen thickness 1 mm).

4 L.V. Raulea et al. / Journal of Materials Processing Technology ) 44±48 47 individual grains. In our case such a ``trend'' could be caused by the recrystallization treatment applied Blanking experiments: effect of large grain sizes Fig. 7. Variation of tension and bending yield strength with thickness-tograin-size ratio j. the reproducibility of the test decreases strongly with increasing grain size, and b) the mean force increases with increasing grain size. The loss of reproducibility is directly related to the reduction in the number of grains that are actually loaded in the deformation zone. Individual grains will vary in orientation, and inevitably lead to large variations in process force, as it is indicated by larger standard deviation values of the bending force. The observed increase of the bending force with increasing grain size could be rationalized similarly. As the single grain behavior is dominated by its orientation, the observed trend of increase of the mean process force may be related to the orientation of the The results of blanking experiments on specimens of 1 mm thickness are shown in Fig. 8. In Fig. 8a, the deformation of a polycrystalline specimen with a grain size of mm 2 is presented. In Fig. 8b and c, two single crystal specimen with different orientations are shown. In all three gures, the punch displacement is the same. The highlighted contours in Fig. 8 show the blanking product shape at this deformation stage. In the case of the polycrystalline specimen Fig. 8a), the upper and lower rollover specimen zones have the same radius. The deformation process is point symmetric with respect to the middle point of a line connecting the punch lower-right radius with the die upper-left radius, as it was also pointed out by Stegeman et al. [4]. Typically, this point symmetry is lost in the case of the single-crystal specimens, as it is shown in Fig. 8b and c. It is also clear that the product shape varies strongly, as illustrated by the difference between Fig. 8b and c. The lack of reproducibility of the product shape is likely to be related to the variation of the crystal orientation. The deformation behavior of single crystals is extremely anisotropic and therefore will depend strongly on the crystal orientation. The geometry of the blanked edge is shown in Fig. 9 for three different single crystals. A strong variation in the size Fig. 8. Different blanking shapes above): a) polycrystalline, b) single-crystal I, and c) single-crystal II specimens of 1 mm thickness, with the displacement grid on the deformation zone. 1) Upper rollover zone and 2) lower rollover zone. Below, the calculated equivalent strain field.

5 48 L.V. Raulea et al. / Journal of Materials Processing Technology ) 44±48 Fig. 9. Blanked edges of three different single crystals: 1) rollover zone, 2) shearing zone, 3) fracture zone specimen thickness 1 mm). Fig. 10. Blanking force versus punch displacement for single crystal specimens and a polycrystalline specimen. of the rollover zone, the smooth-sheared surface, and the fracture zone can be observed. Fig. 10 presents the blanking load of single crystal and polycrystalline specimen as a function of punch displacement. The polycrystalline force curve was found to be reproducible for several specimens, therefore the mean force curve is represented. The blanking force of the single crystal specimen shows a strong variation with respect to: a) the maximum force and b) the curve pro le. These extreme differences can be rationalized by taking into consideration that the deformation during the blanking process takes place in the most favorable grain gliding system. Since the grain orientation varies after recrystallization, the gliding systems are oriented differently with respect to blanking direction. Moreover, after some deformation, including possibly grain rotation, another gliding system could become active and modify the grain deformation behavior. As a result, the single crystal blanking force is not reproducible, and furthermore varies in magnitude and pro le from the polycrystalline blanking force. 4. Conclusion 1. Uniaxial tensile tests were performed on Al 2S in order to investigate the influence of thickness reduction at a constant grain size. The experimental results show a decrease of the yield strength with the decreasing specimen thickness. This effect attributed mainly to the influence of the free specimen surface. 2. Bending experiments were employed on Al 2S in order to study the influence of increasing the grain size at a constant specimen size. The experimental results show an increase of the yield strength with decreasing grain size. This effect attributed to grain-size strengthening or Hall±Petch effect. If the grain size becomes larger than the specimen thickness, the yield strength tends to increase with grain size and a loss of reproducibility is observed. This can be rationalized by the fact that in the deformation zone, only a few grains are loaded. Moreover, the behavior of each grain differs strongly due to variations in orientation. 3. In single crystal, blanking experiments on Al 2S, the loss of reproducibility of product shape and process force are observed. These are caused by the different grain orientations of blanking specimens. The grain orientation and its anisotropy have a dominant role in plastic deformation behavior and are responsible for changes in the development of blanking product geometry. References [1] R. Kals, F. Vollertsen, M. Geiger, Scaling effects in sheet metal forming, in: H.J.J. Kals, et al. Eds.), Proceedings of the Fourth International Conference on Sheet Metal, Vol. II, Enschede, 1996, pp. 65±75. [2] R.T.A. Kals, R. Eckstein, M. Geiger, Miniaturization in metal working, in: Proceedings of the Sixth International Conference on Sheet Metal, Twente, 1998, pp. 15±24. [3] S. Miyazaki, H. Fujita, H. Hiraoka, Effect of specimen size on the flow stress of rod specimens of polycrystalline Cu±Al alloy, Scripta Metall ) 447±449. [4] Y.W. Stegeman, A.M. Goijaerts, D. Brokken, W.A.M. Brekelmans, L.E. Govaert, F.P.T. Baaijens, An experimental and numerical study of a planar blanking process, J. Mater. Proc. Technol ) 266±276. [5] T.H. Courtney, Mechanical Behavior of Materials, McGraw-Hill, New York, 1990.